Is engineering a bubble that will burst?

Just because data is imperfect doesn’t mean it is useless. Imperfect data just means you have to search harder to determine the range of likely outcomes. There is absolutely value in knowing for example that researchers interested in becoming a college professor have an extraordinarily difficult time in doing so, or that the term “starving artists” exists for a reason. Individuals can be highly successful in either of those areas, but the odds are against them.